On Sat, Jan 13, 2007 at 01:26:47PM -0800, Andrew Sackville-West wrote: > On Sat, Jan 13, 2007 at 10:23:00AM -0500, Thomas H. George wrote: > > A binary scan works, a color scan stalls the scanner. > > > > A posting to sane-devel on 1/6/07 drew no responses. > > > > The backend maintainer pointed out that "If the USB system stalls, the > > scanner can not get rid of it's data". I tried scanimage -B to increase > > the buffer size but this did not help. Actually, after reinstalling > > xsane and libsane I was able to complete one color scan but the second > > one stalled. > > try this. get your one good scan, then pull the usb cable (while > watching syslog) and then put it back in. Make sure udev removes and > inserts modules correctly and try again.
Wild result. Pulling scanner cable - no problem. Pulling wacom tablet cable - partial lockup. tablet not restored, xsane can't find scanner, xane cannot be killed, lsusb locks one terminal, can open another terminal but lsusb cannot be killed, rebooted. Since I had to reboot I first downloaded linux-source-2.6.18, built the new kernel and rebooted to it. After this I was able to complete four color scans before the scanner stalled on a fifth scan. I tried deleting everything in /tmp which is the directory I set in xsane. This did not resolve the problem. My assumption: Somewhere there is a buffer that usb uses and it is not being cleared. Even rebooting the system does not resolve the problem though earlier reinstalling xsane temporarily solved the problem - i.e. one color scan worked before the scanner stalled. > Also, one thing that can cause usb > to stall would be too much stuff coming down the pipe. what else do > you have plugged into that usb channel? lsusb would help here. Maybe > you need to split some of your usb stuff onto different controllers. The lsusb output: Bus 004 Device 010: ID 04b8:011b Seiko Epson Corp. Perfection 2400 Photo Bus 004 Device 001: ID 0000:0000 Bus 001 Device 003: ID 056a:0015 Wacom Co., Ltd Bus 001 Device 001: ID 0000:0000 Bus 003 Device 003: ID 051d:0002 American Power Conversion Back-UPS Pro 500/1000/1500 Bus 003 Device 002: ID 03f0:8204 Hewlett-Packard (the printer) Bus 003 Device 001: ID 0000:0000 Bus 002 Device 001: ID 0000:0000 Baring a power failure there should be nothing extra going on during a scan. > > > > > The system is Debian testing with a 2.6.16 kernel. The scanner is an > > Epson Perfection 2400 which is recognized as a usb2 device. It has > > worked flawlessly with xsane for the past several years. Recently I did > > several apt-get updates and apt-get dist-upgrades. It is after these > > upgrades that I have encountered the problem. > > do you know what version? check the changelogs? might need to file a > bug about regression. > > > > > > I have confirmed that the hard drive is using dma and 32 bit i/o. There > > is 775 Mb memory and all the partitions have at least 2 Gb of free space. > > looks good to me. > > .02 > > A -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]